Another reason those Antidepressants might not be working: taken Aspirin lately?

Aug 10 2011 Published by under Behavioral Neuro

A few days ago I talked a little bit about the interactions between stress and neurogenesis, and the way that this might impact symptoms of depression. But it turns out that there’s even more to the way stress affects the body than impacting neurogenesis, and the long term effects could in turn affect the way antidepressants work, or don’t work.

You see, antidepressants aren’t all that effective. This is mostly because we’re not sure how antidepressants work, or even have their effects when they DO work. Until we understand the mechanism, we can’t come to an understanding of what better drugs we can use, and why the current drugs are only partially effective. But there are other effects that stem from our lack of mechanistic understanding. For example, because we don’t understand HOW antidepressants work when they do work, we ALSO don’t understand how antidepressants might interact with other drugs that might increase or decrease their effectiveness.

And here’s where we get to cytokines. Cytokines are chemicals that are specifically involved in inflammation. They are released from the glia in the your central nervous system and can modulate your immune system, promoting or suppressing formation of antibodies, among other things. And inflammation has a link to depressive behavior, though we’re not really sure what it is. Many people who are getting treated with cytokines (drugs to decrease cytokines are used in treatment of autoimmune disorders) will often suffer from depressive symptoms, presumably as a result of the treatment. And some people think of major depressive disorder as a partial dysregulation of sickness behaviors. Slow moving, decreased cognitive abilities, decreased interest in fun activites, these are all behaviors that are not only associated with depression, they are associated with sickness in general. Not only that, several kinds of cytokines can interact with the hypothalamic-pituitary axis, which regulates stress responses and is also involved in depressive behaviors. So it is possible that regulation of cytokines can help regulate the HPA axis, whichi in turn can influence depressive like behavior. In one direction this means that it could make depression worse, but in another, it could also mean that it could make depressive symptoms better.

So how do you figure this out? You look at drugs that decrease inflammation, thus decreasing cytokine activity. The best well know of these are the non-steroidal anti-inflammatory drugs, otherwise known as the NSAIDS, and the best known of THOSE are Aspirin and Ibuprofen. So could NSAIDS, by affecting levels of cytokines, affect the HPA axis and antidepressant responses?


Warner-Schmidt, et al. “Antidepressant effects of selective serotonin reuptake inhibitors (SSRIs) are attenuated by anti-inflammatory drugs in mice and humans” PNAS, 2011.

To look at this question, the authors of this study looked in both a mouse model, as well as in humans, using a huge study known as the STAR*D study (which I'll get to in more detail later). They also looked at mice treated with the antidepressant citalopram (commonly marketed as Celexa) and the NSAID ibuprofen, to see how these two interacted.

They started with the mice. Using chronic treatment with the SSRI Celexa, divided the mice into four groups. Half got Celexa, and half got saline. And half of each of THOSE groups got Ibuprofen.

What they found was that Celexa treatment increased levels of cytokines in the brain, and that ibuprofen treatment BLOCKED the increase in cytokines. You can see along the bottom of the graph a whole list of cytokines. The gray bars are the mice that received just catalpa, and you can see increases in the expression of all the cytokines listen. The clear and lighter gray bars are the mice that received ibuprofen. You can see that ibuprofen blocked the effect of the celexa on the cytokines, and decreased cytokines in normal mice (as you’d expect actually, it’s kind of Ibuprofen’s JOB to do this).

Ok, so Celexa increases cytokines and ibuprofen blocks this increase. But the question now is whether that increase in cytokines when mice are given Celexa is actually behaviorally functional. To look at this, the authors performed a behavioral test for depressive-like behavior in mice while giving them ibuprofen. The idea is that if INCREASES in cytokines help with antidepressant behavioral effects, then a decrease in cytokines produced by ibuprofen should have the opposite effect, producing a more depressive-like effect.

In this case they used the tail suspension test. In this test a mouse is strung up by its tail on a rod. You leave it there for 6 minutes, and see how much time it spends just hanging, as opposed to struggling and trying to get away (though the really hard part is making sure the highly agile mouse doesn’t climb its own tail and ruin your experiment). When left hanging there for a while, the mouse will struggle, but will eventually give up and show something called behavioral despair, where it hangs by its tail and waits for the experimenter to let it down. When you give a mouse antidepressants, however, the mice will wriggle around for longer, and hang less. What you can see here is the white bars, where mice got various kinds of antidepressants, all decreased the amount of time the mice spent hanging. But when the authors gave the animals ibuprofen in ADDITION to the antidepressants, it reduced the effects, causing them to spend more time immobile. So it appears that ibuprofen can indeed decrease the behavioral effects of antidepressants (though it apparently has no behavioral effect on its own).

And of course, you have to see if it works the other way.

In this graph you can see the effects in normal mice in the tail suspension test when they increased levels of cytokines. On the left you can see that the cytokines had antidepressant like effects, reducing the amount of time the mice spent immobile. And on the right, you can see that this effect was blocked when the mice were knockouts of p11, a protein which is induced by cytokines and by antidepressants and may be responsible for some of the effects.

So it looks like NSAIDs may be able to block the behavioral effects of antidepressants in mice by blocking increases in cytokines that the antidepressants induce. That’s all well and good in mice, but what about in humans? In this case the scientists took the data from the STAR*D study: sequenced treatment alternatives to relieve depression. This study looked at a large population of people with and without depression, and with and without antidepressants, in an effort to find which drugs were most effective, and what other variables might impact how people become depressed and respond to antidepressants. In this case, they asked a bunch of patients who were being treated for depression with Celexa for at least three months if they had taken an NSAID at any point in time during those three months. They found that patients who had been treated with Celexa alone showed a 55% success rate of treatment in three months. That’s not very good, but those who had also taken NSAIDs showed only a 44% success rate. This means that the NSAIDs could be impacting the ability of the antidepressants to function, making the Celexa work even worse than it already does.

Of course, there are several issues with this study: in the humans, what are they taking the NSAID FOR??? Chronic pain? That may make them harder to treat. Also, the patients were asked if they had taken an NSAID any time in the last three months. They could have taken Aspirin once and it would have counted. Obviously this is an already conducted human study and you work with what you have, but future studies should certainly be more careful, say limiting the questions to people who took Aspirin daily. Not only that, HOW are the NSAIDs and Celexa interacting? Does this influence things like hippocampal neurogenesis or the function of the HPA axis? And how? Obviously we’ll need to run some more studies before we can tell how much of an effect this is having. But not only does this study show us possible drug interactions between antidepressants and NSAIDs (future studies could show that people with depression need to watch out for aspirin, which is bad news for people in chronic pain), it also gives us one more clue into the mechanisms of how antidepressants work, and how to make them better.

Warner-Schmidt JL, Vanover KE, Chen EY, Marshall JJ, & Greengard P (2011). Antidepressant effects of selective serotonin reuptake inhibitors (SSRIs) are attenuated by antiinflammatory drugs in mice and humans. Proceedings of the National Academy of Sciences of the United States of America, 108 (22), 9262-7 PMID: 21518864

7 responses so far

  • becca says:

    I notice NSAID + bupropion has the opposite effect- it worked (a smidge) better with the NSAID. I really wonder if it has similar effects on the cytokines, since it works so differently. Probably good since it causes killer headaches.

    And for the human population- yeah, it's tough. particularly since all analgesics seem to have a detrimental effect on treatment- seems more likely *pain* makes it harder to recover.

    • Zuska says:

      Are you saying buproprion causes headaches? Because my neurologist says it has some anti-migraine effect.

      Question to self: Do I keep taking the buproprion for the anti-migraine, and anti-depression over chronic pain + other crap, AND Aleve for the incessant neck pain?

      Or is it all pointless?!?!?

      • becca says:

        It definitely gives me headaches (note- we're talking high doses here, e.g. 400mg/daily). But then, they were normal everyday headaches that respond quite nicely to ibuprofen (or even tylenol, which is slightly less effective for me). I've never had a migraine. It was an unpleasant side effect, but hardly the most worrying possible.
        I get the feeling that my headaches : migraines :: sweating during exercise : fever from malaria.

        Keep in mind, the authors speculate explicitly on the interaction being dependent on 'effects in the periphery' because NSAIDS do not generally cross the blood-brain barrier. In other words, the data to date are entirely consistent with some NSAIDs upregulating some drug metabolizing enzyme that results in removal of the drug from your body faster. To my mind, this is actually the simplest explanation (though it does rely on handwavy powers of drug metabolizing enzymes)
        Thus, all else being equal
        1) get the XR (or ER) formulation of buproprion if it doesn't cost you any more
        2) don't take it at the same time as the aleve
        3) don't drink too much grapefruit juice (weird stuff can upregulate your drug metabolizing enzymes, grapefruit juice is a classic example)

        Also, apparently this mysterious p11 that the cytokines are thought to act through is related to BDNF. Which can also reliably be raised by exercise. So, if the NSAID can help you get moving, I suspect the opposition to antidepressant effects will be somewhat circumvented. But that's quite speculative.

  • Seamonkey says:

    What's interesting is that elevated levels of cytokines are found in subjects with MDD, and in some studies decrease with treatment. It would seem like the pain meds, then, should actually improve the symptoms. There is some evidence that there is a difference in HPA axis function in the MDD population, especially atypicals and melancholic. Atypicals trend to have more responsive HPAs and melancholic HPAs are flatter, so having an accidental sampling bias can yield results that seem fly in the face of the cytokine hypothesis of MDD. It's messy, and frustrating me right now!

  • Vicki says:

    I suspect the answer to this is "no," but it seems worth asking: are there any studies about NSAIDs and welbutrin or other non-SSRI antidepressants?

  • cheshire says:

    Reading this I wonder if there is anything known about whether Paracetamol would be a better alternative for anti depressant taking people? Nothing turned up in a quick search of pub med :